The jobs we'll lose to machines -- and the ones we won't | Anthony Goldbloom

Translator: Matias Hall Reviewer: Anders Finn Jørgensen

So this is my niece.

Det her er min niece.

Her name is Yahli.

Hun hedder Yahli.

She is nine months old.

Hun er ni måneder gammel.

Her mum is a doctor, and her dad is a lawyer.

Hendes mor er læge, og hendes far er advokat.

By the time Yahli goes to college,

Når Yahli studere på universitetet

the jobs her parents do are going to look dramatically different.

vil hendes forældres arbejde se meget anderledes ud.

In 2013, researchers at Oxford University did a study on the future of work.

I 2013 lavede forskere fra Oxford en undersøgelse om fremtidens arbejde.

They concluded that almost one in every two jobs have a high risk

De konkluderede, at næsten hver andet job

har stor risiko for at blive automatiseret af maskiner.

of being automated by machines.

Machine learning is the technology

Machine learning er den teknologi

der ligger til grunde for størstedelen af ændringerne.

that's responsible for most of this disruption.

It's the most powerful branch of artificial intelligence.

Det er den mest virkningsfulde del af kunstig intelligens.

It allows machines to learn from data

Det tillader maskiner at lære fra data

and mimic some of the things that humans can do.

og efterligne nogen af de ting, som mennesker kan gøre.

My company, Kaggle, operates on the cutting edge of machine learning.

Mit firma, Kaggle, arbejder på det nyeste inden for machine learning.

We bring together hundreds of thousands of experts

Vi samler flere hundrede tusinder eksperter

to solve important problems for industry and academia.

som skal løse vigtige problemer for den industrielle og akademiske verden

This gives us a unique perspective on what machines can do,

Det giver os et indblik i hvad maskiner kan,

what they can't do

hvad de ikke kan,

and what jobs they might automate or threaten.

og hvilke job de måske kommer til at automatisere

Machine learning started making its way into industry in the early '90s.

Første gang machine learning blev brugt industrielt, var i start 90'erne.

It started with relatively simple tasks.

Det startede simpelt.

It started with things like assessing credit risk from loan applications,

Det startede med at vurdere kreditrisiko fra låneansøgninger,

sorting the mail by reading handwritten characters from zip codes.

brev sortering ved at læse håndskrevne tegn fra postnumre.

Over the past few years, we have made dramatic breakthroughs.

Gennem de sidste par år er der sket banebrydende gennembrud.

Machine learning is now capable of far, far more complex tasks.

Machine learning er nu i stand til langt mere komplekse opgaver.

In 2012, Kaggle challenged its community

I 2012 udfordrede Kaggle sit lokalsamfund

to build an algorithm that could grade high-school essays.

til at programmere en algoritme til at bedømme gymnasie stile.

The winning algorithms were able to match the grades

Algoritmen der vandt, gav den samme karakter

given by human teachers.

som den rigtige lærer gjorde.

Last year, we issued an even more difficult challenge.

Sidste år lavede vi en sværere udfordring.

Can you take images of the eye and diagnose an eye disease

Kan man tage billeder af øjet og diagnostisere en øjensygdom

called diabetic retinopathy?

kaldet diabetisk retinopati?

Again, the winning algorithms were able to match the diagnoses

Igen gav algoritmen der vandt den samme diagnose

given by human ophthalmologists.

som var givet af en øjenlæge.

Now, given the right data, machines are going to outperform humans

Maskiner udkonkurrerer altid mennesker ved opgaver som denne,

at tasks like this.

givet at den får de rigtige data.

A teacher might read 10,000 essays over a 40-year career.

En lærer læser måske 10.000 stile over en 40-årig karriere.

An ophthalmologist might see 50,000 eyes.

En øjenlæge ser måske 50.000 øjne

A machine can read millions of essays or see millions of eyes

En maskine kan læse millioner af stile eller se millioner af øjne

within minutes.

på få minutter.

We have no chance of competing against machines

Vi kan ikke hamle op med maskinerne,

on frequent, high-volume tasks.

når det gælder mængde opgaver.

But there are things we can do that machines can't do.

Men der er ting, vi kan, som maskiner ikke kan.

Where machines have made very little progress

Maskiner er ikke blevet særlig meget bedre

is in tackling novel situations.

til at takle unikke og nye situationer

They can't handle things they haven't seen many times before.

De kan ikke arbejde med ting, de ikke har set en masse gange før.

The fundamental limitations of machine learning

De grundlæggende begrænsninger for machine learning

is that it needs to learn from large volumes of past data.

er at de skal lære fra store mængder tidligere data.

Now, humans don't.

Det skal mennesker ikke.

We have the ability to connect seemingly disparate threads

Vi har evnen til at finde sammenhængen i forskellige situationer,

to solve problems we've never seen before.

og løse problemer vi ikke har set før.

Percy Spencer was a physicist working on radar during World War II,

Mens fysikeren Percy Spenser arbejdede med radarer under 2. verdenskrig,

when he noticed the magnetron was melting his chocolate bar.

opdagede han at magnetronen smeltede hans chokolade bar.

He was able to connect his understanding of electromagnetic radiation

Han forenede sin forståelse for elektromagnetisk stråling

with his knowledge of cooking

med sit kendskab til madlavning

in order to invent -- any guesses? -- the microwave oven.

for at opfinde -- nogen gæt? -- mikrobølgeovnen.

Now, this is a particularly remarkable example of creativity.

Det er et eksempel på særlig kreativitet.

But this sort of cross-pollination happens for each of us in small ways

Den forenende måde at tænke på, sker for as alle på mindre stadier

thousands of times per day.

tusinder af gange om dagen.

Machines cannot compete with us

Maskiner kan ikke hamle op med os,

when it comes to tackling novel situations,

når det kommer til unikke situationer,

and this puts a fundamental limit on the human tasks

og det skaber en begrænsning for hvilke opgaver

that machines will automate.

maskiner komme til at automatisere.

So what does this mean for the future of work?

Så hvilken betydning har det for fremtidens arbejde?

The future state of any single job lies in the answer to a single question:

Om et job er sikret for fremtiden, kan besvares med et spørgsmål:

To what extent is that job reducible to frequent, high-volume tasks,

I hvor stor en grad kan jobbet nedskrives til mængde opgaver,

and to what extent does it involve tackling novel situations?

og i hvor stor grad indebærer det unikke situationer?

On frequent, high-volume tasks, machines are getting smarter and smarter.

Når det gælder mængde opgaver, bliver maskiner klogere og klogere.

Today they grade essays. They diagnose certain diseases.

I dag bedømmer de stile. De diagnosticere visse sygdomme.

Over coming years, they're going to conduct our audits,

De følgende år, vil de foretage vores regnskaber.

and they're going to read boilerplate from legal contracts.

De kommer til at læse standard tekst fra juridiske kontrakter

Accountants and lawyers are still needed.

Revisorer og advokater skal stadig bruges.

They're going to be needed for complex tax structuring,

De skal bruges til kompleks skatte-strukturering,

for pathbreaking litigation.

til banebrydende retstvister.

Men maskiner vil mindske deres omdømme,

But machines will shrink their ranks

and make these jobs harder to come by.

og gøre den type jobs mindre hyppige.

Now, as mentioned,

Som jeg har nævnt,

machines are not making progress on novel situations.

laver maskiner ingen fremskridt når det gælder unikke situationer.

The copy behind a marketing campaign needs to grab consumers' attention.

Rammen for en marketing kampagne skal fange folks opmærksomhed,

It has to stand out from the crowd.

den skal skille sig ud.

Erhvervs strategi går ud på at finde mangler i markedet,

Business strategy means finding gaps in the market,

things that nobody else is doing.

noget som ingen andre gør.

It will be humans that are creating the copy behind our marketing campaigns,

Det kommer til at være mennesker som skaber rammen for en marketing kampagne,

and it will be humans that are developing our business strategy.

og det vil være mennesker, som vil udvikle vores erhvervs strategi.

So Yahli, whatever you decide to do,

Så hvad end du beslutter dig for at lave, Yahli,

let every day bring you a new challenge.

lad hver dag bringe nye udfordringer.

If it does, then you will stay ahead of the machines.

For så vil du forblive foran maskinerne

Thank you.

Tak.

(Applause)

(Klapsalve)